过去 24 小时 10 条报道、8 条推特
2026-11-21 08:08 至 2026-11-22 08:08 (北京时间)
模型
1. OpenAI Ships GPT-Live-1.1 for Voice Agents, Cuts Idle Cost to $0.02/Min
OpenAI has released GPT-Live-1.1 in the API. The full-duplex voice model improves multilingual accents and transcription, and a new standby mode bills idle WebRTC time at $0.02 per minute instead of $0.05. The post OpenA
来源:MarkTechPost
产品
1. How to build a knowledge factory for AI agents
The knowledge factory is how LangChain keeps its agents' company knowledge current: agents flag what's missing, and a curator agent fixes it in the background.
来源:LangChain Blog
行业
1. 8万条群聊消息做成考卷:腾讯元宝×KCL发布首个基于群聊构建的Agent记忆基准
8万条群聊消息做成考卷:腾讯元宝×KCL发布首个基于群聊构建的Agent记忆基准
来源:量子位
2. 【数智周报】 Anthropic敲响纳斯达克开市钟,创史上最大规模IPO纪录;美国限制远程获取AI算力,海外数据中心出租算力须许可;DeepSeek第二轮融资落地
(11 月 16 日-21 日)英伟达史上首次单季营收突破1000亿美元,净利润同比翻倍;阿里巴巴:2027财年第二季度营收2715.2亿元,同比增长10%,AI相关产品收入超179亿元,连续13个季度实现三位数增长;腾讯:WorkBuddy近七天年化收入超5亿美元,三季度收入是第二季度的5倍以上;豆包工作日活突破800万,企业客户ARR突破20亿元...
来源:钛媒体
3. Bytedance argues most AI training phases are unnecessary and backs it up with a competitive open model
Bytedance and several US universities introduce Axion 1.0, an AI agent that rivals top commercial models from OpenAI and Anthropic on benchmarks after just three training phases. The researchers attribute the gains to ca
来源:The Decoder
4. Can Anthropic afford to put safety first?
Today, I’m talking with Hayden Field, The Verge’s senior AI reporter, about the Anthropic IPO for the third and final episode in our series about AI in real life. The AI lab behind Claude went public on the Nasdaq last T
来源:The Verge AI
5. Anthropic mathematician claims he used Claude to prove 19th-century Riemann hypothesis
Anthropic’s Levent Alpöge says it took six weeks for him using the AI tool to solve 19th-century Riemann hypothesis A mathematician at Anthropic has said that the San Francisco -based artificial intelligence (AI) company
来源:Guardian AI
6. ‘I was shaking, crying, having trouble breathing’: the unchartered trauma of breaking up with AI companions
People are falling in love with AI systems like ChatGPT and Gemini - and some relationships are ending in ways that feel like the very human experience of being dumped Sónia Veenstra, a 48-year-old former clinical psycho
来源:Guardian AI
7. OpenAI shares new prompting tips for GPT-6.2 Astra and Sol to fix brief answers and unnecessary questions
OpenAI has released new prompt templates for GPT-6.2 Astra and Sol meant to fix common model behaviors such as overly brief answers, unnecessary follow-up questions, and excessive testing when coding. The API is changing
来源:The Decoder
8. Hours after Anthropic claims a Claude proof of the Riemann hypothesis, Stanford releases an all-human competing draft
Anthropic mathematician Levent Alpöge says he used Claude to prove the Riemann hypothesis, math's most famous open problem for 167 years, with a proof formalized in Lean that spans over 250,000 pages. Just hours later, S
来源:The Decoder:AI News
推特
1. Gemini 4 tip: the difference between Gemini 4 Hydrogen on "high" thinkin…
Gemini 4 tip: the difference between Gemini 4 Hydrogen on "high" thinking effort and that same model on "xhigh" is VERY significant Here's the high v.s. the xhigh pelican - high cost 3.99 cents and used 5,311 output toke
来源:X · Simon Willison
2. Little known fact about the man of the hour Levant @__alpoge__, is that…
Little known fact about the man of the hour Levant @__alpoge__, is that he cooks at Age of Empires 2 Marlo Van Marck: I really recommend reading the statement from Dr. Alpöge on his progress solving the Riemann Hypothesi
来源:X · Dylan Patel
3. BREAKING: Jensen Huang reportedly told NVIDIA employees that if AMD's Mi…
BREAKING: Jensen Huang reportedly told NVIDIA employees that if AMD's Minimax Inferact kernels beat NVIDIA's on AgentX performance targets, he would reset every NVIDIA employee's GTA 6 progress. The employees then broke
来源:X · SemiAnalysis
4. More price competition for open models is good (but hard when GPU supply…
More price competition for open models is good (but hard when GPU supply is so constrained) dax: we are popping the pricing bubble today starting with deepseek flash and glm 5.3 we are pricing 60% lower than fireworks/ba
来源:X · Nathan Lambert
5. Lots of work we've done to make Antigravity the best harness for complex…
Lots of work we've done to make Antigravity the best harness for complex tasks using Gemini. Neon using AGY CLI is now #1 on AA Coding agent Index and a massive jump over mini-swe on SWE-sweep, a bug finding and fixing e
来源:X · Varun Mohan
6. OpenAI models topping the alignment benchmarks: Arena.ai: The Arena Alig…
OpenAI models topping the alignment benchmarks: Arena.ai: The Arena Alignment Index has been updated! We've refined how our signals detect model failures, providing broader coverage and more reliable judgements. Based on
来源:X · Greg Brockman
7. Give Grok @Bot your hardest tasks Molly O’Shea: Grok Bot planned a 5,000…
Give Grok @Bot your hardest tasks Molly O’Shea: Grok Bot planned a 5,000-person conference Ron Biscardi (@rbiscardi) says his team used Grok Bot for literally everything to set up the TomorrowX Summit at record speed: “G
来源:X · Elon Musk
8. Going to check out this threads thing
Going to check out this threads thing
来源:X · Tibo
本篇在北京时间早上 08:08 生成,汇总此前 24 小时的报道和推特,标题直链原文。
